JP6757727B2 - Device-based motion-compensated digital subtraction angiography - Google Patents

Description

The present invention relates to image processing methods, image processing systems, computer program elements, and computer readable media.

Multiple medical interventions are performed under fluoroscopy or angiography. In other words, a real-time projected X-ray image is acquired to image the inside of the patient and a medical device or device is introduced into the patient to realize the intervention.

For example, in order to stop the growth of cancer tissue or AVM (arteriovenous malformation), an embolus is intentionally formed in its feeding artery to close the blood supply (AVM) and / or stop the supply of nutrients to cancer or growth. (Transcatheter arterial chemoembolization (TACE), etc.). This type of intervention, called medical embolization, is performed by administering an embolus-forming agent to a desired location (region of interest (ROI)) in the human body using a catheter tube. The embolic forming agent is essentially a liquid volume or mass of glue containing a small bead suspended in the carrier liquid, thereby occluding the affected area. During such embolization interventions, it is extremely good to ensure that only the feeding arteries of interest are blocked and healthy vessels are not. Currently, the location of the embolus is monitored by acquiring one or more fluoroscopic projection images. Due to the opacity of the embolism and / or the radiation of the carrier fluid, a projected "footprint" can be recognized within the fluoroscopic image, thereby providing the intervention radiologist with clues as to the location of the embolus. ..

Another example of an intervention that relies on image-based support is transcatheter aortic valve transplantation (TAVI). To assess the outcome of TAVI treatment, a visual assessment of valve regurgitation is routinely performed using angiography. Injection of contrast agent is performed. The intervener then visually inspects the angiographic frame for the presence of contrast medium regurgitation into the left ventricle and thus determines the severity of the regurgitation (amount of contrast in the ventricle, filled with contrast). Based on the percentage of ventricles that were created).

In the above two exemplary and other interventions, the contrast within one or more motor layers is inadequate when viewed against a complex and messy background (spine, ribs, medical devices, etc.). Often the image quality is impaired. Digital Subtraction Angiography (DSA) may be used to improve this situation and achieve better image contrast. In conventional DSAs, the mask image is subtracted from the contrast image (eg, angiography or fluoroscopy frame that captures the embolus). However, due to the complexity of the movement patterns during the intervention, the DSA may introduce subtractive artifacts, which also impair image quality.

Therefore, alternative image processing methods and / or related devices are needed in the art to address at least some of the defects mentioned above.

An object of the present invention is solved by the subject matter of the independent claims, and further embodiments are incorporated in the dependent claims. It should be noted that the following aspects of the invention apply equally to image processing methods, image processing systems, computer program elements, and computer readable media.

According to the first aspect of the present invention
An input port configured to receive at least one mask image of at least some test pieces containing an object and at least two projected images including at least one contrast image, with different acquisition times for the mask image and the contrast image. The input port, which is acquired in and represents the region of interest (ROI) with different contrasts,
A target discriminator configured to identify at least one target of the object in a contrast image and at least one mask image.
A motion estimator configured to estimate the motion of an object by the motion of at least one identified target over at least a contrast image and at least one mask image, wherein the motion is a motion estimate with respect to the motion of the ROI. With a vessel
A motion compensator configured to align at least one mask image with at least one contrast image based solely on the estimated motion of the target.
A subtractor configured to subtract at least one aligned mask image from at least one contrast image to obtain a difference image of ROI.
An image processing system is provided with an output port configured to output the difference image.

According to one embodiment, the image processing system comprises a visualizer configured to display at least a portion of the difference image corresponding to the ROI on the display device.

In other words, the proposed method implements a ROI motion compensation DSA focused on the ROI motion. More specifically, the movement of the ROI is obtained by tracking a target located inside or outside the image portion representing the ROI, and only the movement of this target is considered to compensate for the movement of the ROI. To. In other words, motion compensation is based solely on the motion of the target. Other movements of surrounding image objects are ignored. Therefore, an object that undergoes a movement different from that of the target may appear blurry in some cases, but this blurring allows the operator to (in some cases) be more prominent in the ROI against a blurred background. It has been discovered by the applicant that this is actually an advantage, as it makes it easier to focus mentally on the high contrast image information of. This is when the proposed method aims to visualize images in real time to assist demanding interventions where the operator must successfully navigate complex anatomical structures and movement patterns. , Especially advantageous.

According to one embodiment, the identification of the target is based on auxiliary image data aligned with at least one of the projected images.

According to one embodiment, the target receives the first movement, and the alignment motion of the motion compensator corresponds to the position of the target by the selected mask image corresponding to the position of the target by at least one contrast image. Includes selecting a mask image to do so.

According to one embodiment, the image processing system comprises a target designation input port, and the classifier identifies the target in response to receiving designation of one or more target at the input port. This designation is a selection from a mask image or a contrast image.

According to one embodiment, the visualizer is configured to display at least a portion of the difference image along with the mask image and / or contrast image on the display device. In one embodiment, only one section of the difference image is displayed, i.e. this section is "ROI concentrated".

According to one embodiment, the visualizer is configured to display at least a portion of the aligned auxiliary image data.

According to one embodiment, the system is configured to process at least two projected images with respect to the second target and / or the second ROI to obtain a second difference image, and the visualizer is the difference. It is configured to display at least a portion of the second difference image in place of the image or together with the difference image.

According to one embodiment, the visualizer is configured to display a graphical overlay in the mask image indicating the location of the ROI and / or the second ROI.

According to one embodiment, the target receives a second movement and the motion compensator
i) The position of the target for the first movement with the two additional mask projection images will be substantially the same, and ii) the position difference for the second movement with the mask image and at least one contrast image will be the target. Two additional mask images are configured to be selected so that they are substantially the same with respect to the object.

According to one embodiment, the subtractor
The two additional mask images are subtracted to obtain a mask difference image, and after motion compensation for the first or second movement, the mask difference image is subtracted from the difference image to acquire a cascade difference image. Will be done.

According to one embodiment, the visualizer is configured to display at least a portion of the cascade difference image on the screen.

According to one embodiment, the target is, or relates to, a projected footprint on a pristine or external object, particularly an implanted object such as a prosthetic heart valve or embolus. The object is present in the test piece at each acquisition time of the mask projection image and the projection image.

An exemplary embodiment of the present invention will be described below with reference to the following drawings.

It is a figure which shows the imaging arrangement. It is a flow chart of the image processing method by 1st Embodiment. It is a flow chart of the image processing method by 2nd Embodiment. It is a figure which shows the graphics display by one Embodiment. It is a figure which shows the exemplary image produced by the image processing method proposed in this specification.

With reference to FIG. 1, the basic components of a fluoroscopy or angiography imaging arrangement that can be used to assist intervention procedures are shown.

The heart valve of the patient SP is not functioning normally. During the TAVI intervention procedure, the healthcare professional introduces a guide wire into the femoral artery of the patient SP and then guides the delivery catheter OB to the affected aortic valve ROI to be repaired or replaced. As the guidewire advances through the patient's cardiovascular structure P, a series of continuous fluoroscopic images F are acquired by the X-ray imager 100. Another example is an embolization procedure in which a catheter OB for embolization agent administration is guided to the site of AVM or cancer (similar to TACE).

During the intervention, the patient SP is placed on bed B between the X-ray tube XT of the X-ray imager 100 and the detector D. The X-ray tube XT and the detector D are attached to a rigid frame C rotatably mounted on the bearing B. Fluorescence fluoroscopic image operation is controlled from the computer console CC. The intervention radiologist can control image acquisition via the console CC and can "photograph" a series of individual fluoroscopic frames ("fluoro") F by activating a joystick or pedal. .. According to one embodiment, the imager 100 is a C-arm type, but other systems are also envisioned.

During image acquisition, X-ray radiation exits the X-ray tube XT, passes through the ROI, is attenuated by interaction with substances in the body, and then the beam p thus attenuated constitutes detector D. It hits the surface of the detector D in one of the plurality of detector cells. Each cell hit by the beam reacts by issuing a corresponding electrical signal. The group of signals is then converted into their respective digital values representing the attenuation. The density of the materials that make up the ROI determines the level of attenuation, with higher densities causing greater attenuation than lower densities. The digital values so registered for each X-ray p are then aggregated into an array of digital values forming a fluoroframe for a given acquisition time and projection direction. In other words, each fluoro is a digital image of a projection along the projection direction, the direction of which is determined by the rotation of the C-arm at a given acquisition time or moment. The series of fluoroFs is then digitally processed by the data acquisition unit DAS and then sent to the image processor IPS. Its purpose and operation will be described in more detail below.

In one embodiment, the detector D is a component of an image enhancement tube that projects an image directly onto the screen M for real-time observation.

In the fluoroscopic image F, it is generally possible to recognize the footprint of only a significantly attenuated object. More specifically, only guide wires or catheter OBs made of high opacity material appear as projection footprints or "shadows" within each fluoroF. The flow of the fluoroscopic image F is a frame rate sufficient to monitor the progress of the guidewire by the catheter OB through the body of the patient SP (about 15 images per second for TAVI, or about 15 images per second for TACE). 2 images per second).

In some interventions, a high opacity contrast agent (“dye”) is used to increase the contrast to the parenchyma, or to image the dynamic behavior of the fluid, as in the assessment of valve regurgitation. Delivered to SP. A certain amount of contrast agent (“bolus”) then travels along with the bloodstream through the vascular structure and finally to the ROI. In other words, the dye or contrast agent imparts temporal opacity to the normally invisible vascular structure ROI, and a vascular tree appears as a spider-like footprint within each vascular AG. On the way to the ROI, the bolus acquires a fluoroscopic frame of ROI and a higher contrast fluoroframe at the point where the bolus passes through the ROI. These specially timed fluorographs are angiographic diagrams. In other words, the flow of the fluorescence fluoroscopic frame F comprises two types of frames, one acquired in the absence of contrast medium in the ROI (this is the default, but in this context, the inventor We call these "dye-free" frames "mask image" MI), and some frames are acquired during the presence of contrast agent in the ROI. These are angiographic frames or "contrast image" AGs. An embolus-forming agent may be used instead of the contrast agent, and therefore the above description regarding the timing of the mask image and the contrast image applies. The same term "contrast image" is used regardless of whether a contrast or embolic agent is used or any other material is used.

The imaging system guides each medical device OB (catheter, valve to be implanted, etc.) to the ROI, as well as the actual in the ROI such as embolization, heart valve repair, or any other diagnostic or evaluation procedure. Intervention tasks can also be supported. For example, in an embolization intervention, this task releases a volume of embolization agent (hereinafter referred to as "glue mass", "embolization", or simply "mass") to the ROI through the catheter system. That is. The ROI is, for example, a shunt of blood vessels that needs to be occluded because the patient needs to be treated for AVM, arteriovenous fistula (AVF), or hemangiomas. Examples of liquid embolizing materials include Onyx® (a glue-like substance), alcohol, or n-butyl cyanoacrylate (NBCA). Administration of emboli, by releasing the volume of embolization agents around ROI through the open end of the catheter, starting at the moment t _0. The embolus then circulates in the bloodstream and is then held at the target location (usually the shunt that connects the arterial and venous systems), thereby occluding the blood vessels.

Other interventional procedures include interventions for cardiovascular repair, such as assessment of regurgitation in a natural or prosthetic heart valve. In this case, the ROI is considered to be the location of the implanted prosthesis.

DSA is used to further increase image contrast within angiography AG and eliminate the contribution of attenuation from background or foreground objects and other clutter, with DSA traditionally presenting angiography. In the present specification, the mask image MI (fluoro in which only zero or negligible contrast medium is present in ROI) is subtracted from the contrast image (also referred to as contrast image) to create a differential image DF1 (in pixels). Unfortunately, however, patient movements, such as physiological activity (breathing and / or heart), often cause movements during the acquisition of fluoro / angiography. This has the effect that the image object (eg, representing the ROI and the respective projected footprint of the catheter or other instrument OB and / or organ, such as the ribs) undergoes obvious movement over the frame. Simply subtracting frames from each other when movement occurs can lead to the introduction of artifacts.

In particular to better assist images of the above or similar interventions in motion, the system further comprises an image processor configured to act on the flow of fluoroscopic images provided by an X-ray imager. FIG. 1A to be inserted shows the details of the image processing system IPS proposed in the present specification.

The IPS includes an input port IN and an output port OUT. There is a target classifier module LID, a motion estimator ME, and a motion compensator MC. These modules process frames with different flow of fluoroscopic images and pass over the processed frames to the subtractor module, as described in more detail below with reference to FIGS. 2 and 3. Proceed to DIF to create the differential image DIF1 and therefore perform DSA. The difference image DIF1 is then output at the output port OUT and then stored for further processing or proceeds to the visualizer VIS, which interacts with the video equipment of the imaging system 100 to display unit M ( A graphics display GD is created on the monitor, screen, etc.), and then the display unit M displays the difference image DIF or a part thereof.

Simply put, the proposed image processor system acts as a contrast enhancer to increase the image contrast only in the portion of the image that represents the ROI. The image processor IPS proposed herein acts essentially as a locally motion-compensating DSA module. The present invention proposes an optimized subtraction that focuses only on the ROI around a unique target in the image. The target may also be located outside the ROI as long as there is a known deterministic relationship between the movement of the target and the ROI. The target is an external object, such as the tip of a catheter OB, or any other pre-introduced / implanted device such as a heart valve, or the partially embolized tissue or the embolus itself. It may be. A dedicated algorithm can process the image and track the target within a later image. The more structural the features the target provides, the easier and more robust it is to detect across frames. The IPS then performs a subtraction that accurately moves and compensates only for the local area of interest (generally, not always the area around the target). Therefore, the rest of the image may have significant subtraction artifacts. However, with some interventions, artifacts in image parts that do not represent ROI are considered okay because the information in those image parts is of little importance. More specifically, the image processor IPS operates to compensate for the movement within one or more layers of motion that the object being considered receives. A "moving layer" is one in which some image objects appear to move over different frames, while other image objects, such as background image objects, appear to be stationary or moving in different ways. , Refers to the phenomenon in the projected image. Therefore, it can be said that these objects are located in different moving layers. The methods of the invention can handle multiple targets within the region of interest, but they are processed separately to create their own motion-compensated differential image DFI.

The operation of the image processor IPS will be described in more detail below with reference to the flow charts of FIGS. 2 and 3 showing different embodiments.

With reference to the flow chart of FIG. 2, FIG. 2 describes the operation of the image processing system IPS for the motion compensation DSA for a single motion of the target OB to be considered. For example, this is by no means limited and is for illustration purposes only, but the first movement is induced by cardiac activity and the second movement is induced or driven by respiratory activity. Yes, it is suppressed by using the respiratory arrest protocol assumed herein. FIG. 3 below describes different embodiments that can handle combined movements. Again, movements other than heart or respiration, especially non-periodic movements, are considered herein by development and are specifically envisioned herein.

In step S210, the mask image frame MI and the contrast image frame (angiography frame) AG are taken out from the flow F of the fluoroscopic image. This can be achieved, for example, by using a thresholding method that monitors the gray level intensity at individual pixels within each frame of the fluoroscopic flow. Then, for example, it can be determined whether the currently received fluoroscopic frame is an angiographic view AG.

In steps S220 and S230, motion analysis is performed on these two or more image frames AG, MI. More specifically, in step S220, the target (eg, footprint or shadow of the image) of the introduced object OB is detected or identified within the two image MIs and AGs. Examples of suitable targets are the tip of the catheter (eg, through which the embolus is administered), the shape and absorption signature of the embolus itself currently deposited in the area of interest, or the implanted device, eg. The shape signature of the heart valve for the application of cardiac surgery can be mentioned. However, pristine objects such as ribs, calcifications, or other identifiable organs are also considered suitable targets. Also note that the target is not always within the area of interest, but in some cases this is the case. All that is needed is that there is an a priori known deterministic relationship between the movement of the area of interest and the movement of the observed target. In other words, in some embodiments, the movement of the target is considered a surrogate for the movement of the ROI. After the movement of the target is detected, this known dynamic relationship allows the associated movement of the region of interest to be calculated. The target identification step 220 is essentially a tracking operation across the two images MI and AG. If multiple targets are identified per image, each target is tracked or "tracked" separately. After the target is identified across the two images, flow control proceeds to step S230, from which the movement of each object OB is estimated from the targets identified across the two images.

Flow control then proceeds to step S240, where essentially alignment with respect to the estimated motion is performed. More specifically, the proposed alignment relates only to the area of interest and / or the movement of the target to be considered. According to one embodiment, other movements or other parts of the image are not considered. This enables motion compensation that concentrates on the ROI that is characteristic of the proposed image processor IPS. The ROI concentrated motion compensation scheme of step S240 includes a rigid or non-rigid alignment scheme such as warpage that allows two targets to be aligned with each other over the two image MIs, AGs. Then, in order to essentially align the two images and thereby compensate for the considered movement within the motion layer due to the movement of the target, one of the considered images MI, AG is modified, eg, shifted or It is transformed in other ways. In one embodiment, multiple targets for one ROI are considered to acquire non-rigid alignment. In the case of non-rigidity, deformations (local motion descriptors) can be constructed to effectively attempt morphing of a given mask image and contrast image pair against each other. However, in some stiffness alignment embodiments, the motion descriptor is as simple as a single shift / transformation vector or simple rotation.

In an alternative embodiment, in the alignment step of S240, the recorded motion phase of the target (or the position of the target if the motion is not periodic) corresponds to the motion phase of the target by the current contrast image AG. As such, it comprises a selection step of targeting and selecting a mask image from a plurality of previously received mask images. For example, a set of mask image frames can be acquired before the arrival of the contrast bolus in the region of interest to be considered. Thereby, a "stockpile" of mask images is acquired, and each mask image captures the target in a different cardiac phase. These mask images are recorded throughout the cardiac cycle. Then, in the alignment step, the position / configuration of the target in the later acquired contrast image is compared with the position of each target in the previously recorded stockpile of mask images to the current target by contrast image AG. Take out the one that best corresponds to the position of. For example, this selection is achieved by overlaying the current contrast image on each of the buffered masks MI to establish where the greatest overlap between targets occurs. For example, each position of the implanted heart valve footprint is used in the context of this heart. However, this is also useful in the case of embolus formation, as it is known to change shape fairly slowly as the bolus hardens, and therefore the heart or breathing or any other movement of interest that one wants to compensate for. Tracking is possible with a sufficiently high frame rate.

In other words, in one embodiment, the latest (ie, recent) mask image can be retrieved from the fluoroscopy flow F, which is then obtained by deformation based on the detected movement of the target. It is aligned on the current contrast image. In other embodiments where the alignment is performed by selection, the mask image is selected because the position of the captured target is selected so that it is essentially the same as the position of the target in the current contrast image. The mask image does not necessarily have to be the latest mask image and may be an older frame from the flow F. Therefore, you have to go back a few frames to find the best mask. However, even when so selected, there may still be residual inconsistencies in some movement of the target by the two images, which will be corrected by a suitable pixel shift algorithm in step S245 of arbitrary selection. be able to. However, in one embodiment, there is no such optional step and the alignment is from a previously recorded mask image to the "appropriate image", i.e. the current contrast image with respect to the position of the target. Relies only on the alignment achieved by selecting the image that best corresponds to the position of the target by. Optional step S245 includes, instead of, or in addition to, pixel shifting, a filtering operation that filters a portion of the target. For example, at least the footprint of the target in the mask image can be completely smoothed or removed to further prevent artifacts in the differential image DFI. It will be appreciated that in the present specification, the selection step for ensuring the correspondence of the motion phases of the targets recorded in the two images is performed entirely on the image information. In other words, the present specification can completely eliminate the gating technique that requires the use of an external device such as an ECG. However, that does not mean that the proposed method cannot be used with gating signals, and such combinations are also envisioned herein in alternative embodiments.

Then, the two images so aligned in step S240 are sent to step S250, and the current contrast image is subtracted from the mask image MI to acquire the difference image DF1.

Then, in step S260, the difference image is output and can be used for further processing such as display performed in step S270. The various display modes will be described in more detail below with reference to FIG.

With reference to the flow diagram of FIG. 3 below, FIG. 3 shows a DSA algorithm similar to that of FIG. 2, but the embodiment of FIG. 3 is from a combination of various movements or cycles acting on the region of interest / target. It is adapted to handle the movements that occur. Compensation for these combined motion patterns is realized by a double subtractive method.

Simply put, in the embodiment of FIG. 3, the first subtraction is performed in step S350b to compensate for, for example, respiratory movements, and the second subtraction step S350a is performed in the second flow to remain. It is suggested to eliminate the movement of the heart.

More specifically, as in the method of FIG. 2, in step S210, the mask image MI and the contrast image AG are received. Similarly to the above, the motion analysis is executed in steps S220 and S230, and steps S320 and S330 correspond to the above steps S220 and S230.

Based on the analyzed movement, in step S340B, as in step S240 above, as a selection or search operation, the two images are aligned with each other and the received mask image is stored in multiple previously stored (eg, buffered). It is an image so selected from the ringed mask image. However, unlike the alignment step S240 of FIG. 2 above, in step S340b, the selection is here focused only on the respiratory movements. This can be done in motion analysis steps S320, S330 by tracking targets that are more significantly affected by respiratory movements such as ribs. For these targets, the motion components of the heart are negligible. Alternatively, a temporal analysis can be performed to thus identify the two motion cycles via a unique frequency signature.

After the optional filtering step S345B for removing residual motion when alignment is attempted by selection, step S350B forms the first difference image DF1.

In parallel with or following these steps, in step S340A, two other mask images MI2, MI3, each recording a target in a similar respiratory phase, are so selected, but In addition to this movement, the movement resulting from the second movement cycle (particularly the activity of the heart) is also considered. More specifically, in step S340b, the two mask image reference MI2 and MI3 have their cardiac phase difference in the cardiac phase difference between the other two images, that is, in the current contrast image AG and the mask image MI1. Selected to resemble the phase difference (if any).

As in step S345B, also in optional step S345A, as described above in FIG. 2, the filter and / or residual motion correction is applied.

Next, in step S350A, a second difference image DF2 between the two mask reference frames MI3 and MI2 is formed to reach the second difference DF2.

In step S360, the two differential images are aligned with each other, eg, by using the position of the target within the two differential images, eg, the position of the implanted device, eg, the heart valve. Then, in step S370, the two motion-compensated / aligned difference images DF1 and DF2 are subtracted from each other to form a "difference difference" or cascade difference image DDF, and in step S380, the cascade difference image DDF It is output and can be used for further image processing or display in step S390. This will be further described later in FIG.

Although the steps of FIGS. 2 and 3 have been described with reference to periodic movements in some cases, this is only according to one embodiment, and the proposed method can also be applied to non-periodic movements. It should be understood that the specification also assumes.

The embodiment of FIG. 3 has been described with respect to a motion pattern formed from two motions (ie, caused by each driver), which can be achieved by repeatedly applying the embodiment of FIG. 3 to three or more. It can also be expanded into movements. That is, it starts with two movements and continues as described. Then, for example, in order to pay attention to the third motion component, the fourth and fifth mask images are selected in step S340A, and the process is continued accordingly.

The methods of FIGS. 3 and 4 have been described with reference to one target and one ROI, which can also be developed to handle multiple targets, and the images are respectively. The movement of the target is processed separately and independently.

Optionally, the above method according to FIG. 2 or 3 also includes an additional masking step based on 3D segmentation of the region of interest to further enhance visibility / contrast in the lower portion of the region of interest considered. You can also do it. For example, in an exemplary embodiment of the assessment of reflux in the left ventricle, this is done by aligning the valve based on 3D segmentation so that everything outside the left ventricle is masked. Further masking can be realized. Obviously, this double masking can also be applied in applications other than cardiac surgery.

With reference to FIG. 4, FIG. 4 shows various display modes for one or more differential images. According to some embodiments, the entire difference image is displayed on the monitor M. In one particular advantageous embodiment shown in FIG. 4, only the true subordinate portion or "clipping" of the difference image is displayed within the dedicated display area 410AB on the screen and the current fluorescence fluoroscopy frame. Is displayed within 405. In other words, only the ROI portion of the difference image is displayed in the display area dedicated to ROI.

The region of interest itself is indicated by the respective graphical overlays on 415a, b overlaid on the current frame 405, contouring the respective image portion for each ROI. The current frame 405 is a mask image or a current contrast image AG, or this is a conventional DSA without motion compensation. In the latter case, it can be considered that a large display area 405 provides a wide DSA display and one or more smaller display areas 410a, b each provide a local ROI concentrated motion compensation DSA display. .. In other words, within each display area, one section of each difference image DF is displayed, which section is defined by the position and size of each ROI representing the graphical overlay 415a or 415b.

As shown in FIG. 4A, the two image portions contoured by overlays 415a, b for different local ROIs overlap. The differential images for each individual ROI represented by the graphical overlays 415a, b are calculated separately for each ROI, as far as the movement is concerned, taking into account only their own movements by each target. Will be done. As will be appreciated, for each ROI represented by overlays 415a, b, different targets are used to capture the movement of each ROI. For example, in the case of ROI overlay 415a, the target is the footprint of the stent, and in the case of ROI overlay 415b, the target is the footprint of the tip of the catheter.

In one embodiment, the locally motion-compensated ROI image is displayed in an enlarged form within each display area 410AB.

In other embodiments, it is advantageous to show only the region of interest and therefore not the frame 405 currently underneath.

In one embodiment, additional auxiliary image sources such as CT, MRI, or other 3D images can be used to assist in the selection of 3D targets. The image source is then aligned onto the projected image MI, AG, and then the 3D target is projected onto the projected image, thus defining the footprint of the 3D target, thereby allowing the motion compensation operation described above. To. However, the auxiliary image does not necessarily have to be 3D, and in all embodiments, for example, a 2D auxiliary image to assist in the selection of a target is also envisioned. In such an example in which the auxiliary image is used in this way, the corresponding image information from the auxiliary image is graphically placed on the lower section of the difference image displayed in the respective dedicated DSA display areas 410a and b. Overlaid as an overlay. In addition, or instead, the auxiliary image may be displayed as an overlay at each ROI position within the current overview frame 405.

The display mode described above with respect to FIG. 4 can be used for any of the methods according to FIGS. 3 and 2 described above. In other words, the display areas 410a and 410b indicate a section of the difference image according to FIG. 2 or the double difference image according to FIG. Although FIG. 4 shows two local display areas, it will be appreciated that the display modes described above can be used for any number of display areas, including a single local display area.

In terms of assisting user interaction, the proposed system also includes the ability to automatically detect suitable targets, which the user can then choose from. In this way, the target can be seen across all images, thus ensuring that no external image needs to be used. In addition, or as an alternative, the user can select the desired target and / or outline the ROI overlay within an auxiliary image, such as within a 3D CT image volume, and then these , Mapped onto projected images MI, AG, based on the projection direction used by the projection imager 100.

In one embodiment, when the method of FIG. 2 or 3 is applied to a plurality of targets, the target displayed for the user to choose from has the smallest overall alignment error. Selected by the system. In other words, in this embodiment where the target is automatically proposed, the system pre-calculates the respective alignment errors for a given contrast image and mask image pair in the background. More specifically, the selection of the number of targets for independent motion compensation (within a certain range and with a certain penalty when this number increases), position, and size selection should minimize the total alignment error. Is automatically realized. This total alignment error can be quantified by using the respective individual vector fields established in the respective ROI estimates in steps 220 and 240. In one embodiment, this error quantization is achieved by clustering the error maps of a high density local motion estimation technique (such as an optical flow or block matching algorithm) before finding the optimum. However, it is understood that even in this embodiment where common alignment is considered, the actual motion compensation for each ROI is still limited to consider only each motion for the target associated with that ROI. I want to. In other words, motion compensation for each DSA motion is performed independently of each other and thus each differential image is optimized for only one unique ROI / target motion.

Further, in terms of user interaction, the proposed method includes a drag-and-drop method in which the user selects each ROI and / or target by mouse click or touch screen operation and then swipes it on the touch screen. Drag across the screen to each display area by operation or mouse click and drag operation. The target so designated is then sent into the IPS through the target transmit input port LD-IN. This user interaction then triggers the calculations according to FIGS. 2 and 3 above. In another embodiment, the IPS is fully automatic, i.e., the target is automatically identified, selected and processed by FIGS. 2 and 3 without any user input. As an alternative, a manual embodiment is also envisioned, where the automatically selected ROI / target is not proposed to the user and the user is free to enter anything.

In one embodiment, the user separately selects the region of interest itself, eg, by drawing a suitable neighborhood from the points of the circle or square shown in FIG. 4, or by drawing any other suitable geometry. Can be done. This defines the size or "reach distance" of the ROI as the image neighborhood. In some cases, if the target is located inside the contoured ROI, the target is also automatically defined. If the target is located outside the ROI, this target is separate to associate each ROI with each target that should use that movement to achieve motion-compensated DSA operation for each ROI. By enabling the operation of, it can be logically linked to the ROI. For example, a double-click method is used to define the ROI and the target separately, but in pairs with predefined synchrony in the user interaction step. For example, the user first defines the ROI, for example by drawing a square that defines the ROI. The user then clicks on a position outside the ROI and the system interprets this second position as a target to be associated with just the contoured ROI for the purpose of surrogate motion estimation. This association feature between the ROI and the target can be implemented by a suitable counter and event handler scheme.

The above-mentioned operations of FIGS. 2 and 3 and the display operation of FIG. 4 can be performed in real time, that is, while the image frame in the fluoroscopic flow is received at each port IN. Will be understood. However, an offline mode is also envisioned, and each compensation operation according to FIG. 2 or 3 can be performed on previously acquired images, or can be performed in review mode. Although the above method has been described with specific examples in terms of cardiac interventions such as aortic regurgitation assessment or embolization procedures, the present application should be successfully used in any intervention procedure requiring digital subtraction movement. Please understand that you can. Also, the application is not limited to supporting interventional surgical procedures. Also envisioned herein are applications that rely on other images in other areas, such as geology or radar image research.

With reference to FIG. 5, this shows an exemplary image output by the method described above in comparison to a conventional method. These images show the shadow of each footprint / radiation of the prosthetic heart valve after introduction. Pain A shows a conventional angiographic diagram, Pain B shows a conventional subtraction of an angiographic diagram using the latest available mask frames, and Pain C shows a prosthetic heart valve as a target. The result by the method 2 is shown. There are fewer artifacts, especially near the aortic valve. At this time, regurgitation in the ventricles (visible under the transplanted valve) can be more easily evaluated.

The image processing module IPS is arranged as a software module or routine having an interface suitable for reading in the fluoro stream F and is executed on a general purpose computing unit or a dedicated computing unit. For example, the processor IPS runs on the workstation or console CC of the imaging system 100. The image processing module IPS, along with some or all of its components, resides on an executive agency (such as a general purpose computer, workstation, or console) or is remoted by the executive agency via a suitable communication network within a distributed architecture. / Accessed in the center.

Alternatively, the components of the image processing module IPS are arranged as a dedicated FPGA (Field Programmable Gate Array) or similar stand-alone chip. These components are programmed within a suitable scientific computing platform such as Matlab® or Simulink® and then converted into C ++ or C routines maintained in the library for general purpose computers, workstations. Linked as required by an executive agency such as a station or console.

Another exemplary embodiment of the invention provides a computer program or computer program element that is adapted to perform the method steps of the method according to one of the aforementioned embodiments on a suitable system. Will be done.

Therefore, computer program elements are stored on a computer unit, which is also part of an embodiment of the present invention. This computational unit is adapted to perform or induce the steps of the method described above. In addition, this computing unit is adapted to operate the components of the device described above. Computational units can be adapted to operate automatically and / or execute user instructions. The computer program is loaded into the working memory of the data processor. Therefore, the data processor is equipped to implement the methods of the invention.

This exemplary embodiment of the invention includes both a computer program that uses the invention from the beginning and a computer program that updates an existing program into a program that uses the invention.

In addition, computer program elements can provide all the steps necessary to meet the procedures of exemplary embodiments of the methods described above.

According to a further exemplary embodiment of the invention, a computer readable medium such as a CD-ROM is presented, the computer program element is stored on the computer readable medium, and the computer program element is as described above. ..

Computer programs are stored and / or distributed on suitable media such as optical storage media or solid state media supplied with or as part of other hardware, but on the Internet or other wired or other wired or It may be distributed in other forms via a wireless telecommunication system or the like.

However, computer programs are also presented via networks such as the World Wide Web and can be downloaded from such networks into the working memory of the data processor. A further exemplary embodiment of the invention provides a medium that makes a computer program element available for download, such that the computer program element performs a method according to one of the aforementioned embodiments of the present invention. Placed in.

It should be noted that embodiments of the present invention have been described with reference to various subjects. In particular, some embodiments have been described with reference to method type claims, and other embodiments have been described with reference to device type claims. However, from the above and below description, unless otherwise indicated, any combination of features belonging to one type of subject, as well as any combination of features relating to different subjects, will be disclosed by this application. Those skilled in the art will understand that it is considered. However, all features can be combined to provide more than a simple sum of features.

Although the present invention has been illustrated and described in detail in the drawings and the above description, such illustrations and descriptions should be considered descriptive or exemplary and not restrictive. The present invention is not limited to the disclosed embodiments. Other modifications to the disclosed embodiments will be understood and implemented by those skilled in the art by reading the drawings, the present disclosure, and the dependent claims in carrying out the claimed invention.

In the claims, the word "comprising" does not exclude other elements or steps, and the indefinite article "a" or "an" does not exclude plurals. A single processor or other unit may fulfill the functions of some of the items described in the claims. The fact that specific measures are described in different dependent claims does not indicate that a combination of these measures cannot be used in an advantageous manner. Any reference code within the scope of the claims should not be construed as limiting the scope.

Claims (16)

An input port that receives at least one mask image of at least some test pieces containing an object and at least two projected images including at least one contrast image, wherein the mask image and the contrast image are acquired at different acquisition times. , The input port, which represents the region of interest (ROI) with different contrasts,
A target classifier that identifies at least one target of the object in the contrast image and the at least one mask image.
A motion estimator that estimates the movement of an object by the identified movement of the at least one target over the at least contrast image and the at least one mask image, wherein the movement of the target is the movement of the ROI. With regard to motion estimator and
A motion compensator that aligns the at least one mask image with the at least one contrast image based solely on the estimated motion of the target.
A subtractor that obtains a difference image of the ROI by subtracting the aligned at least one mask image from the at least one contrast image.
An image processing system including an output port for outputting the difference image. The image processing system according to claim 1, further comprising a visualizer that displays at least a part of the difference image corresponding to the ROI on a display device. The image processing system according to claim 2, wherein the identification of the target object is based on auxiliary image data aligned with at least one of the projected images. The target receives the first movement, and the alignment operation of the motion compensator is such that the position of the target by the selected mask image corresponds to the position of the target by the at least one contrast image. The image processing system according to claim 2, further comprising selecting the mask image. The image processing system includes a target designation input port, and the target identifyr responds to receiving one or more designations of the target at the target designation input port. The image processing system according to any one of claims 2 to 4, wherein the designation is a selection from the mask image or the contrast image. The image processing system according to claim 2, wherein the visualizer displays at least a part of the difference image together with the mask image and / or the contrast image on the display device. The image processing system according to claim 3, wherein the visualizer displays at least a part of the aligned auxiliary image data. The image processing system processes at least two of the projected images with respect to the second target and / or the second ROI to obtain a second differential image, and the visualizer replaces the differential image. The image processing system according to claim 2, wherein at least a part of the second difference image is displayed together with the difference image. The image processing system according to claim 8, wherein the visualizer displays a graphical overlay indicating the position of the ROI and / or the second ROI in the mask image. The target receives a combination of two movements, and the movement compensator
i) The position of the target due to the first movement with the two additional mask images will be substantially the same, and ii) the position difference with respect to the second movement will be with the mask image and with the at least one contrast image. The image processing system of claim 4, wherein the two additional mask images are selected so that they are substantially the same as those for the target. The subtractor
Subtract the two additional mask images to obtain a mask difference image.
The image processing system according to claim 10, wherein after motion compensation for the first motion, the mask difference image is subtracted from the difference image to acquire a cascade difference image. The image processing system according to claim 11, wherein the visualizer displays at least a part of the cascade difference image on a screen. The target is a projection footprint on a pristine or external object, particularly an implanted object, present in the test piece at each acquisition time of the mask image and the contrast image. Or the image processing system according to any one of claims 1 to 12, relating to such a projection footprint. A method of operating an image processing system including an input port, a target classifier, a motion estimator, a motion compensator, a subtractor, and an output port.
The input port is a step of receiving at least one mask projection image of at least some test pieces including an object and at least two projection images including at least one contrast image, wherein the mask projection image and the contrast image are Steps and steps that are acquired at different acquisition times and represent the region of interest (ROI) with different contrasts.
A step in which the target classifier identifies at least one target object of the object from the contrast image and at least one mask image.
The motion estimator is a step of estimating the motion of an object due to the motion of the identified at least one target over at least the contrast image and the at least one mask image, and the motion of the target is the ROI. Steps and steps related to the movement of
A step in which the motion compensator aligns the at least one mask projection image with the at least one contrast image based solely on the estimated motion of the target.
A step in which the subtractor subtracts the at least one mask image aligned from the at least one contrast image in order to obtain a difference image of the ROI.
A method in which the output port comprises a step of outputting the difference image. A computer program for controlling an image processing system according to any one of claims 1 to 13, which performs the steps of the method according to claim 14 when executed by a processing unit. A computer-readable medium storing the computer program according to claim 15.

Images